​

Global Options

Global options are shared by multiple Orchestor CLI data commands. Availability depends on the command; check its --help output. Local configuration commands list their supported options on their own reference pages.

Use --help to display usage and supported options for a command.

Use --workspace to select a workspace by ID. You can also set the ORCHESTOR_WORKSPACE_ID environment variable.

Use --json for compact JSON or --pretty for indented JSON. Both wrap the response in an envelope.

Use --format to select json, table, yaml, csv, or raw. The default is human-readable output.

Use --field to extract a value from the response. Dotted paths are supported. --fields is an alias.

Use --raw to print extracted values without JSON quotes or an envelope. Array values are printed one per line for use in other commands.

Use --output to write the result to a file instead of standard output.

Use --no-pager to disable the interactive pager. The pager is disabled by default when standard output is not a terminal.

Use --dry-run to preview the request without sending it to the API.

Use --yes to skip confirmation prompts for destructive operations. This option does not grant permissions.

Use --page-size on paginated list commands to set the maximum number of items per page. It maps to the API’s limit parameter.

Use --page-all on paginated list commands to stream every page as NDJSON, one record per line. Do not combine it with --json, --pretty, or another output format.

Use --stdin to read a JSON resource definition from standard input when creating or updating a resource. --from-stdin is an alias.

Use --timing to print request duration to standard error. Timing stays separate from the data on standard output, so it can be used with JSON output.

Use --readonly to block any request that can change data before it is sent. ORCHESTOR_READONLY=1 does the same. Only reads, and POST requests marked read-only such as sheets values get, are sent. Use it when handing the CLI to an agent that should only look.

Use --wrap-untrusted to wrap external text such as cell values, notes, and titles in <<<EXTERNAL_UNTRUSTED_CONTENT id="…">>> markers, neutralizing model special tokens and spoofed markers. The output also carries an externalContent note. Use it when an AI model will read fetched content, so the content is treated as data rather than instructions.